1st capture test: Is it supposed to look THIS terrible?

[KenInCa]

I recently purchased a usb capture device and tried it for the first time today (see setup details below). The captured video looks AWFUL:
https://vimeo.com/278955216

I'm not going to try to describe all the problems, because I don't know the correct terminology, and a video is worth at least 1000 words.

While capturing, I had a composite video cable going to a small TV as a monitor device. The video on the TV (a very old, cheap at the time, tube TV) looked GREAT, not showing any of the problems I was seeing on my computer monitor.

Looking at the info shown in VirtualDub, I see there were 0 frames dropped, 3 frames inserted, and a current Sync error of 5 ms.

I've attached a snapshot of the timing graph, in case that's helpful in some way.

So what does it all mean? Is this just perfectly normal, and correctable after capture? Or is there a problem with my setup somewhere?

Thanks for reading, and hopefully steering me in the right direction!

Ken

Hardware:
JVC HR-S9600U (my own, bought new in the '90s)
Diamond VC500 USB capture stick (bought new from Amazon)
VC500 connected to JVC via S-VHS cable (and 2 RCA audio cables)

Software:
Win 10 Pro, 64-bit
VirtualDub 1.9.11 (32-bit), no filters installed, YUY2 color space, 720x480 NTSC_M
Lagarith codec 1.3.27.0 (hopefully 32-bit, but no choice was given)

[BarryTheCrab]

The experts here (not me) will perhaps suggest using the S-video connections rather than RCA, plus external TBC and/or an ES-10/ES-15 to combat the tearing at the top.
Welcome to the rabbit hole.

[KenInCa]

I did connect via S-Video cable (it was the TV that I connected with a composite cable). I erroneously called it an S-VHS cable in my OP above. I used to know the terminology, but gads that was like 20 years ago!

I'd looked at getting an external TBD, but I can't justify the cost. This is the first I've heard of come across an ES-10 / ES-15 reference. Based on some quick googling, I assume that's referring to a Panasonic DVD player, DMR ES-1x? Well, turns out my DVD player (/recorder) *is* a DMR-ES15. So I guess I need to research what it can do for me, and how to hook it up (I suppose it must take the VCR's output, and then pass it through to the USB stick). I had no idea it had video pass-thru, nor do I know what it does to clean up the signal on the way out. More to learn, oh joy!!

Rabbit hole indeed! Thanks for your help.

[KenInCa]

I hooked up my DMR-ES15 between the VCR and VC500 capture device and it made a HUGE difference!! It's not 'perfect', as there are still some overscan lines (I think that's what they're called), but I assume I can get rid of them in post.

But I started wondering why the heck my HR-S9600U didn't do a better job. And then it hit me -- the 'TBC' light was out. I had to press the button to turn it on. DUH. I then compared captures of the same segment with the VCR alone, the VCR with TBC off feeding the ES15, and the VCR with TBC on going through the ES15. I could not see a meaningful difference among the three captures. I even looked at several frames and compared them by having the windows perfectly overlapping and flipping among them. Could not see a difference.

So I think I'll leave the ES15 out of the loop, and just make sure I turn on the JVC's TBC.

Thanks, Barry!

[sanlyn]

Glad to see you figured out the tbc business.

Note that no one here can offer detailed restoration or filtering advice for videos that have been posted to the 'net and deinterlaced/re-encoded/degraded by parties using quick-n-dirty methodsover which we have no control. If you want details about post-capture processing, please post original samples in the captured colorspace without re-encoding, and post them here in the forum. Most members will not go to outside download sites for video samples. Also consider that once your internet video is removed, your thread is mostly useless.

Prepare a sample from a capture by editing a video in VirtualDub. Once you have selected the short edited sample, click "Video...", then click "Direct stream copy," then save your edit as lossless avi.

You will use less CPU power during capture by using Huffyuv instead of Lagarith. Lagarith is normally used for lossless intermediate working files in YUY2, YV12, or RGB.

Learn to use the VDub capture histogram to set proper incoming signal levels for capture. Doing this has been explained in dozens of posts and is discussed in several paragraphs in http://www.digitalfaq.com/forum/vide...-settings.html: The histogram image you posted shows that your capture has crushed blacks, a very low black level, and suppressed mids and brights. Sorry to see you're stuck with Windows 10, but many readers here can likely offer help with W10 video processing issues.

Don't mix 32-bit and 64-bit filters. It won't work. Use 32-bit all the way.

Your posted video looks unusually soft for a 9600U player. Was the original tape recorded at extended play (6 hours or more?).

Quote:

Originally Posted by KenInCa

there are still some overscan lines (I think that's what they're called), but I assume I can get rid of them in post.

Video sources don't have overscan. The term applies to displays, not to tapes.

The noise at the bottom of the frame is called head switching noise. It's cleaned up during post processing.

[KenInCa]

Thank you, sanlyn. I hadn't realized that Vimeo does that, but of course it makes sense that they would, just like Youtube. And I also hadn't thought about the fact that rendering the clip from Photoshop would obviously involve some compression and degredation as well. If I need to post another clip in the future, I will certainly do as you suggest. I was clueless how to even do that much, so your instructions are definitely helpful. Of course, in this case, I just needed something close to what I was seeing for a quick answer, and happily Barry led me to that.

There is such a huge learning curve! I read something, somewhere on this forum that led me to choose Lagarith over Huffyuv. I'm using a fairly powerful CPU (i7-6700K, 4GHz), and CPU usage seemed to be 50% or lower during capture the times I looked, but I'll keep that in mind in case I start to see dropouts or CPU usage approaching 100%.

I noticed the red clipping area, but for some reason stupidly attributed it to the wavy, black vertical areas on the sides. But now that I think about it, those would have resulted in the red area being to the extreme left side only, not a completely filled in area from 0 to 15. So more to learn!

The original tape is a commercial tape (that I don't care about) that I bought over 30 years ago. I wanted to use it as a test tape, so as to not add any wear to the tapes I actually do care about. I just looked and see it was created at SP speed. It might just be the age of the tape, or perhaps my VCR is not in the greatest shape.

Thanks so much for your help and advice; I appreciate it very much. Besides learning more about capturing, I still need to figure out how to do post cleanup, and what format I want for the final result. So my days will be filled for a LONG time!

OH! You added even MORE useful stuff while I was replying! 'Head switching noise' -- now that I know what's it's called, I have a good shot of being able to fix it.

[sanlyn]

Quote:

Originally Posted by KenInCa

There is such a huge learning curve! I read something, somewhere on this forum that led me to choose Lagarith over Huffyuv. I'm using a fairly powerful CPU (i7-6700K, 4GHz), and CPU usage seemed to be 50% or lower during capture the times I looked, but I'll keep that in mind in case I start to see dropouts or CPU usage approaching 100%.

If you have a fast system you likely won't have CPU usage problems with Lagarith for capture. If you do have problems it won't be dropouts, which are a tape or playback defect. Any problems with Lagarith taking up process resources during capture would be be dropped/inserted frames and/or audio sync glitch.

Quote:

Originally Posted by KenInCa

I noticed the red clipping area, but for some reason stupidly attributed it to the wavy, black vertical areas on the sides. But now that I think about it, those would have resulted in the red area being to the extreme left side only, not a completely filled in area from 0 to 15.

If you leave black borders in the image (they are most often y=0 black) they will usually appear as a small sharp "spike" at the far left of histograms.
Samples of the way adjusted levels would look in the capture histogram are shown and discussed in the updated capture guide in the post at http://www.digitalfaq.com/forum/vide...html#post45238. A discussion and images of how to use temporary cropping for measuring histograms more accurately and manipulating proc amp controls is shown in the post at http://www.digitalfaq.com/forum/vide...html#post45239.

Quote:

Originally Posted by KenInCa

So more to learn!

It is at first but you'll find that catching on speeds up once you start capturing and and asking questions, LOL.
Keep in mind that everyone who came here knew little or nothing at first about this stuff.

Quote:

Originally Posted by KenInCa

Thanks so much for your help and advice; I appreciate it very much. Besides learning more about capturing, I still need to figure out how to do post cleanup, and what format I want for the final result. So my days will be filled for a LONG time!

Time will pass more quickly than you think, although there are some speed bumps at first. Try browsing through the restoration section for a while to get an idea of how problems are solved. Things have been a bit slow for two weeks or so because of vacations and whatnot, but going back a month or more brings up some real nightmare captures (some of them make yours look better) and a few spirited debates. The oldest archives go back many years.